Dimensional Insight AI

AI Governed by Your Source of Truth.

Dimensional Insight brings AI into the Diver Platform with a single goal: give your team faster, more confident answers from data they already trust. That means Nora, our virtual assistant — and predictive and analytical capabilities built into the platform itself. Focused, responsible help that respects your access rules and answers from your data, not the open internet.
Talk with our AI team
What it does for you

Less time finding answers. More time acting on them.

AI inside the Diver Platform helps your people work faster, see more, and act with more confidence — without leaving the analytics environment they already trust.

Help your people get oriented faster.

New users, occasional users, and frontline people get answers in the moment instead of waiting on a colleague or filing a support ticket. “Dive without diving” — even for the users who never quite learned to dive.

Get the story behind the numbers, not just the numbers.

Summaries of what the page is showing. Suggestions for where to look next. The kind of analysis your team usually does by hand, surfaced as they work.

Catch what you’d otherwise miss.

Outliers and anomalies surfaced as the data changes — at volumes a person can’t reasonably scan, so the right exceptions reach the right people without a manual review of every row.

Get out ahead of what’s coming.

Forecasts and projections help teams plan ahead — for demand, attrition, supply, capacity — instead of reacting after the change shows up in the rear-view mirror.

Bring the rest of your business into the answer.

Policies, training materials, plans, product information, meeting notes — alongside the data. “Suppliers look at Diver once a month and forget how to use it” — Nora is there when they come back.

Industry fit

Built for the way your industry already works.

Dimensional Insight’s AI is shaped by the industries we know best — where the data is complex, the rules are real, and the cost of a wrong answer is high.

Healthcare

Hospital operations, clinical and quality analytics, supply and pharmacy planning, workforce and capacity questions. AI shows up as a forecast for the coming week’s acute patient days, an outlier rating that catches a shift in case mix, a quality measure explained to a unit director without a help-desk ticket. The Diver Platform has been the system of record for healthcare analytics at major health systems for years; AI extends that footing into the daily work of the people who run those operations.

Beverage alcohol

Distributor, supplier, and wholesaler workflows — inventory and demand planning, depletion and shipment analysis, territory and program performance, supplier collaboration. AI shows up as a forecast that informs what to order ahead of the season, an outlier rating on a brand whose performance is slipping, a question from a supplier answered without an export to Excel. The same governed analytics and AI patterns, applied to the vocabulary and rhythms of the trade.

Outlier detection

Walk in on Monday knowing where to look.

Outlier detection flags the values in your data that don’t fit the pattern — at volumes a person can’t reasonably scan. Your team starts the week with the few items that need attention, not with two hours of finding them.
Walk into the meeting knowing where to start.
Outlier detection / assisted analytics view
Outlier ratings surface what’s worth investigating. Values that don’t fit the pattern are flagged from 1 to 9 within each dimension, so analysts see the highest-priority exceptions without scanning every row.
Forecasting

See what’s coming, with the confidence to act on it.

Forecasting projects demand, capacity, attrition, supply, and other forward-looking measures — with confidence bands and accuracy scores, so your team knows how much weight to put on the number before they make the call.
Walk into the meeting knowing where to start.
Forecasting view with confidence bands and accuracy metrics
Forecasts come with their accuracy. Predicted values appear alongside upper and lower bounds, with standard accuracy measures (MASE, RMSE) so teams can judge how much to trust the projection.
Both outlier detection and forecasting run inside your existing Diver Platform environment. No GPUs, no external cloud ML service, no special AI hardware — and the data stays where it already lives.
“Machine learning works best when it’s iterative and paired with stakeholder expertise.”
Predictive AI lives in the data. Nora lives in the portal where people work with it.
The assistant

Get the answer where you’re already looking.

Nora is our AI assistant. She works from the same governed data and page context your team already sees — so answers stay grounded in your business, not in generic AI output.

Help your team get answers without bothering colleagues.

Nora explains what a KPI means and how it’s calculated, points users to the right report or page, and answers the practical questions that come up in the moment — “Where’s this data coming from? How do I share it?”

Walk into the meeting prepared, not still piecing it together.

Nora summarizes what the analysis is showing, suggests where to look next, and helps your team get past the table and into what matters — “What changed? What am I going to get asked about this?”

Connect the data to the rest of what your business knows.

Nora can use approved business knowledge alongside the data — policies, training materials, strategic plans, meeting notes, product information — and approved external sources where they help. Answers reflect how your organization actually works.

Walk into the meeting knowing where to start.
Walk into the meeting knowing where to start.
A sales rep asks Nora how to prep for tomorrow’s Howell visit. Nora directs them to the right page, names what to focus on, and offers a four-point checklist.
The numbers come with the story.
The numbers come with the story.
A user filters to one customer and asks for the analysis. Nora returns a structured diagnosis — what’s improving, what’s slipping, and which brands are driving the gap.
Our approach

Three questions we ask before any AI feature ships:

  1. 1How does this benefit the customer?
  2. 2How do we know we’re getting good answers?
  3. 3How do we know where the data is going?

The four principles below are how we answer them.

Four principles for AI where truth matters.

In healthcare, beverage analytics, and other industries where decisions carry consequences and data carries rules, AI has to answer to your data — not the other way around.
1

Your data is the source of truth.

AI reasons on your Diver Platform data, your definitions, and your business logic — not on whatever the model absorbed from the open internet.

2

The AI sees only what the user is allowed to see.

Existing access controls apply at query time. Every user gets answers from inside their own governed view.

3

The AI knows what it’s looking at.

KPI definitions, calculations, filters, and on-screen context travel with every question. Arithmetic runs on the Diver Platform’s calculation engine, not the language model.

4

You choose the model. You can change it tomorrow.

Bring your preferred LLM — cloud, BAA-backed, or on-prem. No lock-in, no re-architecture.

Control & privacy

Your environment. Your rules. Your model.

Organizations are not all in the same place on AI. Some want a straightforward cloud connection. Others need a Business Associate Agreement with the model provider, or a fully self-hosted model that keeps every prompt and response inside their own network. Dimensional Insight supports the range.
Nora always runs on top of an LLM that you choose — OpenAI, Anthropic, or any compatible option. You keep direct control of the provider relationship, credentials, and billing.
If pricing, terms, or compliance needs change, you change a setting. Not an architecture.
Customer-selected LLMsBAA-backed cloud optionsSelf-hosted model optionsRole-based visibilityGoverned access controls
Getting started

Talk with our AI team.

Most engagements start with a short conversation — where AI would meaningfully help, and what the right first step looks like for your environment.
01

Identify the right use cases

Where AI will earn its keep — and where it shouldn’t be used at all.

02

Prioritize likely ROI

Sequence the work by where value shows up first.

03

Prove the value

Run a focused proof of concept on a real use case in your environment.

04

Implement where it matters most

Roll out in the workflows your teams already trust.

Schedule a discovery call

Frequently asked questions

Readers expand only what they need.
What is Nora?
Nora is Dimensional Insight’s AI assistant. Users access her through the configurable web portal they already use for dashboards, reports, and analytics, then ask questions in plain English. She works from page context, approved metrics and metadata, supporting knowledge sources, and user-specific access rules.
What can Nora do today?
Explain KPIs, find reports and documentation, summarize and interpret what’s on the page, retrieve approved internal knowledge, and — when enabled — take next-step actions inside the Diver Platform.
What predictive analytics capabilities are available?
Forecasting, outlier detection, classification, regression, and clustering — applied to demand, capacity, attrition, supply, and other forward-looking measures. Models run on the standard Diver Platform infrastructure, with the data staying inside your environment.
How is privacy and security handled?
Nora operates within existing access boundaries. She only sees what the current user is allowed to see. Administrators can configure prompts and context at the portal, environment, and page levels.
Does this support HIPAA and BAA requirements?
Yes. Deployment options include cloud-provider approaches backed by a Business Associate Agreement, as well as fully self-hosted LLM options.
Can customers choose the AI model?
Yes. Customers can connect OpenAI, Anthropic, and OpenAI-compatible endpoints. They maintain direct control over the provider relationship, credentials, and billing.
Can Nora work with more than dashboard data?
Yes. Nora can use both structured analytics data and approved organizational knowledge, including policies, training materials, strategic plans, meeting notes, product information, and documentation.
Can Nora take action, or does she only answer questions?
Both. With AI Actions enabled, Nora can execute scripts, trigger workflows, update records, and report back on the result.
What version of the Diver Platform is required?
The in-product experience on customer data requires Diver Platform 8.0. Version 7.2 supports AI-enhanced documentation, which can serve as an earlier entry point.
What orchestration tooling supports the knowledge-base capability?
For knowledge-base and advanced workflow scenarios, Dimensional Insight supports an orchestration layer and has recommended an open-source option called Dify.ai. Customers are not locked into a single tooling choice.
What vector database options are used for retrieval?
For retrieval-augmented generation and knowledge-base workflows, examples include Weaviate and Elasticsearch.
What does self-hosted architecture mean here?
Organizations can connect the Diver Platform to a self-hosted LLM, and the knowledge-base orchestration layer can also be hosted within the customer’s own environment.
Do we need extra infrastructure?
Nora itself adds minimal overhead to the portal environment. Predictive analytics runs on the standard Diver Platform servers you already have. The knowledge-base capability requires dedicated infrastructure rather than sharing production workloads.
What VM sizing is suggested for the knowledge-base server?
Suggested sizing is 4 cores / 8 GB RAM for up to 100 users; 4–8 cores / 16 GB RAM for up to 200 users; 8 cores / 16–32 GB RAM for up to 500 users.
What is available now, and what is on the roadmap?
Available now: Nora, structured and metadata retrieval, unstructured knowledge bases, predictive analytics, AI-enhanced documentation, customer choice of LLMs, privacy options, and implementation support. Roadmap items include AI updates for beverage applications and a possible Dimensional Insight Cloud-hosted private LLM.
How does an engagement typically begin?
A short discovery conversation to identify useful opportunities, low-hanging fruit, and likely ROI — then a focused proof of concept before any broader rollout.

See what targeted, governed AI looks like inside the Diver Platform.

If you are evaluating how AI can support decision-making without compromising governance, privacy, or control, we’d be glad to show you what that looks like in practice.
Schedule a discovery call